Glossier Data Biography
This data biography and dataset were made for a Data Storytelling class. For this project, Excel was used for dataset creation, and Pivot Tables were used to generate insights with the data. Flourish was used to create visualizations for insights, to better visualize the data story. Below is the data biography and link to view the dataset used with this project.
Data Biography
Origin
All data came from Glossier’s official Instagram, where likes and comments are public data.
The data collection period was between August 22nd, 2022 and October 12th, 2022.
Engagement rate was calculated directly in Excel using standard Excel formulas. Pivot tables were used to generate insights.
Spreadsheet
Click here to view the spreadsheet
Each row represents one post from Glossier’s Instagram account. The date signifies the post (for example, the 22-Aug-22 post represents the Instagram post that was made on that day). The spreadsheet has 52 rows total.
The spreadsheet has a total of 10 columns. Each column contains a variable for the post. For example, there is a column for the day of the week a post was posted, if the post was a giveaway or product announcement, and more.
Transform
All data was manually entered, due to it being collected via individual posts. Calculations were made for engagement, engagement rate, and engagement rate per 100,000.
To calculate engagement, the formula =d2+e2 was used. For engagement rate calculation, the formula =I2/2600000 (engagement divided by amount of followers). To calculate engagement rate per 100,000, the formula =j3*100000 was used.
Analysis
The main function for analysis used was pivot tables, doing univariate and multivariate analyses of the variables to gather insights.
Data was placed into Flourish to create visualizations of the data.
Insights
My insights were:
Glossier’s Grid photos did better than other post types in both likes and engagement.
Posts made on Thursday had the highest average likes and engagement rate per 100,000.
Monday’s had the lowest average likes and lowest engagement rate per 100,000.
Giveaways and product announcements seem to influence post engagement.
Giveaways and product announcements seem to influence post engagement.
Audiences seem to respond more to product photos, rather than photos of human subjects.
Grid photos with human subjects got the highest average likes.
Glossier mostly posts grid photos, with products as the main subject.
My anecdotes were:
As a student, the days I’m primarily active on social media after classes and during the weekends. Other people working on different schedules will be active during other times–for those who work night shifts, they may be less active during the day as they sleep.
Glossier has utilized giveaways for engagement. As shown in the post, Glossier required people to follow, like and comment on the post to count as an “entry.” This carousel received over 4,700 comments, significantly more than their other posts.
Visualizations:
13 visualizations total were made, using Flourish.
Three displayed comparisons between grid posts, carousels and reels with their respective average engagement per 100,000, average comments and average likes. Bar graphs were used.
One displayed the average likes, comments and engagement per 100,000 posts received on each day of the week. A grouped bar graph was used.
Three displayed average likes, comments, and engagement rate per 100,000 for posts that were giveaways versus posts that were regular posts. Grouped bar graphs were used.
Three displayed average likes, comments and engagement rate per 100,000 for posts that contained a product subject versus a human subject. Grouped bar graphs were used.
Two compared the average likes, comments and engagement rate per 100,000 for each post type and content type (one visualization for product subjects and one for human subjects, with respective averages for each post type.)
One is a pie chart showing the percentage of posts with a human subject versus a product/promotional subject.